De novo Mutation Enables NOTCH3 ECD Aggregation and Mitochondrial Dysfunction via Interactions with BAX and BCL-2

Clinical assessment of participants and genetic analysis

Two unrelated probands from different families in China were suspected to with CADASIL disease (genealogical tree presented in Fig. 1a). Informed consent was obtained and peripheral venous blood samples from the probands and their family members were collected for NOTCH3 mutation analysis. A total of 120 healthy participants were treated as controls. Written informed consent forms were provided by all participants before being included in the study. The study was authorized by the Ethics Committee of the Henan Provincial People’s Hospital. The new mutation was affirmed using Sanger sequencing and checked with online databases such as gnomAD (https://gnomad.broadinstitute.org), ExAC (http://exac.hms.harvard.edu/), and HGMD (http://www.hgmd.org/). The skin of the left upper limb of two probands were cut to a depth of about 1.5 cm×1 cm and fixed in 2.5% glutaraldehyde phosphate buffer solution. The fixed skin tissue specimens were sent to the Department of Pathology (Xiangya Hospital, Central South University) for GOM detection.

OPEN IN VIEWER

Fig. 1

Clinical assessment of participants and bioinformatic analysis. a) Pedigrees of two families. The square represents a man; the circle represents a woman. The arrow refers to the proband. Clinically affected family members are represented in black. Unknown clinical status is indicated by a question mark. Deceased members are represented by single diagonal lines. Gene-tested participants are marked with an asterisk. The three diagonal lines indicate carriers. b) Gene sequencing diagrams. Wild type (Normal) and c.1690G > A (p. A564T) mutation (Patients). c) Representative images of skin biopsies of two probands. GOM deposits are revealed under electron microscopy (red arrows). d) Representative MRI in affected family members. A-B), white matter lesion involving the temporal pole and LI in proband of F1; C-E), white matter lesions involving external capsule, LI and magnetic sensitivity weighted imaging showing microhemorrhage in proband of F2; F), perivascular space in basal ganglia in III.2 of F2. e) The Alanine(Ala) at position 564 in NOTCH3 (red square) is highly conservative among Human (NP_000426.2), Chimpanzee (XP_003316212.1), Bovine (XP_003586294.1), Rat (NP_064472.2), Mouse (NP_032742.1), Western clawed frog (NP_001011355.2), and Zebrafish (NP_571624.2). f) 3D model of NOTCH3 is built from Ala to Thr at position 564 in NOTCH3. WT: wild type, MUT: p. A564T mutation.

We used the algorithms of VarCards (http://varcards.biols.ac.cn/), Mutation Taster (http://www.mutationtaster.org/ChrPos.html), PolyPhen2 (http://genetics.bwh.harvard.edu/pph2/), and Combined Annotation Dependent Depletion (https://cadd.gs.washington.edu/) to prefigure whether this mutation effected NOTCH3 function. Subsequently, the homologous amino acid sequences of this mutation and other species were retrieved using Uniprot (https://www.uniprot.org/). DNAMAN software and WebLogo (https://weblogo.berkeley.edu/logo.cgi) were used to analyze evolutionary conservation of the amino acids by aligning sequences of NOTCH3 orthologues of multiple species. Furthermore, we used SWISS-MODEL (https://swissmodel.expasy.org/) to construct pre-mutant and post-mutant protein models, whereas the models of wild type (WT) and mutant were simulated by the PyMol software.

Stably transduced cell lines

Human embryonic kidney (HEK) 293T cells (Shanghai, China) were cultivated in Dulbecco’s modified Eagle’s medium (DMEM) (Hyclone, Logan, UT) added 10% fetal bovine serum (FBS) (Gibco) at 37°C in 5% CO2. The lentiviral vectors expressing WT and A564T mutant of NOTCH3^ECD and a void vector were purchased from BrainVTA Co., LTD (Wuhan, China). A total of 3×10⁴ 293T cells/well were prepared in a 24-wells plate. The following day, the cells in each well were transduced with lentivirus at a multiplicity of infection (MOI) of 6. After 24 h, the transduction media was removed and replaced with fresh DMEM added 10% FBS. Subsequently, the transduced cells were selected using the antibiotic puromycin (2μg/mL) (Sigma, USA) and passaged once every three days in a ratio of 1:3 (Supplementary Figure 1).

Mitochondrial protein extraction

The cells (5×10⁶) harvested were centrifuged at 500 x g for 10 min at 4°C and washed using 0.9% sodium chloride solution. The cell pellet was resuspended with 1 mL ice-cold Lysis Buffer added the Protease Inhibitor Solution (Qiagen, Germany) and hatched for 10 min at 4°C in an end-over-end shaker. After centrifugating at 1000 x g at 4°C for 10 min, the cell pellet resuspended in 1.5 mL of ice-cold Disruption Buffer added the Protease Inhibitor Solution (Qiagen, Germany). The cells were mechanically disrupted using a syringe. After centrifugating again at 1000 x g at 4°C for 10 min, the supernatant collected was centrifuged at 6000 x g for 10 min at 4°C. The pellet contained the mitochondrial fraction and was resuspended in a Disruption Buffer containing the Protease Inhibitor Solutionand. Quantitatively analysis was performed.

Western blot

Cells were harvested and lysed in immunol precipitation (IP) lysis buffer (Beyotime, Shanghai, China), and an ethylenediaminetetraacetic acid (EDTA)-free complete protease inhibitor cocktail tablet (Thermo, USA) was added to extract the total proteins. The mitochondrial proteins were obtained using the Qproteome Mitochondria Isolation Kit (Qiagen, Germany) as described above. Proteins were quantitatively analyzed by the Pierce^TM BCA Protein Assay Kit (Thermo, USA) and separated through sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene fluoride membranes (Millipo, USA). The membranes were blocked in 5% non-fat milk for 1 h and hatched overnight at 4°C with primary antibodies including anti-NOTCH3 mouse monoclonal antibodies (1:1000; Novus, USA), anti-cleaved caspase-3 (1:1000; Abcam, Cambridge, UK), anti-caspase-8 antibodies (1:1000; Proteintect, Wuhan, China), anti-BCL2 (1:1000; Proteintect, Wuhan, China), anti-BAX rabbit polyclonal antibody (1:1000; Proteintect, Wuhan, China), anti-cytochrome c (1:1000; Proteintect, Wuhan, China), anti-VDAC1 (1:2000; Proteintect, Wuhan, China), and anti-β-actin (1:2000; Beyotime, Beijing, China). Protein products were analyzed by the Western Bright™ ECL Spray using horseradish peroxidase (HRP)-labelled secondary antibodies (Millipo, USA). The expression levels were normalized to β-actin/VDAC1 levels in corresponding samples that were quantified on a parallel western blot, wherever indicated. Data were analyzed by densitometry using the ImageJ software.

RNA extraction and quantitative reverse transcription-polymerase chain reaction (qRT-PCR) analysis

The RNA simple Total RNA Kit (TIANGEN) was used to extract total ribonucleic acid (RNA) from the transfected cells. The primers were as follows: β-actin (forward, 5^′ ATCATGTTTGAGACCTTCAACA 3^′ and reverse, 5^′ CATCTCTTGCTCGAAGTCCA 3^′) and NOTCH3 (forward, 5^′ GCGGGTGGGCAGTGTGTGGAT 3^′ and reverse, 5^′ CGTCCACGTCGTCCTCACAGTTATCAC 3^′). Data were obtained by performing three independent in vitro experiments.

Immunocytochemistry

Stably transduced HEK 293T cells (5×10⁴) were plated in 24-well plates containing poly-L-lysine-coated coverslips. They were fixed using 4% paraformaldehyde for 30 min. After washing thrice with phosphate-buffered saline (PBS), the cells were permeabilized using 0.3% Triton X-100 for 30 min. Nonspecific binding was blocked using 5% bovine serum albumin (BSA) in PBS containing Tween detergent® (BSA-PBST) for 30 min. The cells were incubated with anti-NOTCH3 mouse monoclonal antibodies (H00004854-M02, Novus, USA, 1:200 diluted) overnight at 4°C. After washing twice in 1% BSA-PBST, the cells were incubated with Alexa488-labeled goat anti-mouse secondary antibody (Invitrogen, USA, 1:200 diluted) or Alexa564-labeled goat anti-mouse secondary antibody (Invitrogen, USA,1:200 diluted) for 1 h. Nuclei were dyed with 5μg/ml 2-(4-Amidinophenyl)-6-indolecarbamidine dihydrochloride (DAPI) (Beyotime, Shanghai, China) for 5 min and washed twice with PBS for 5 min to remove any excess dye. Subsequently, the cells on the poly-L-lysine-coated coverslips were mounted using a drop of the Antifade Mounting Medium (Beyotime, Shanghai, China). Subsequently, the cells were imaged on the Olympus Model BX53 fluorescent microscope (Olympus, Tokyo, Japan). For co-staining with Thioflavin S (Thio-S) and NOTCH3 antibody, following fixing with 4% paraformaldehyde and washing with PBS, the cells were stained with Thio-S (1 mg/mL), diluted 1:50 in PBS, for 5 min. The remaining steps are the same as above. Data were obtained by performing three independent in vitro experiments.

Cell counting kit-8 (CCK8) assay

The cells (8×10⁴) were seeded in 96-well plates. After 24 h, 10μL of cell counting kit-8 solution (Dojindo, Japan) was added, and the cells were incubated for 4 h. The absorbance value at 450 nm was detected with a microplate spectrophotometer (Synergy H1, BioTek Instruments, USA). Data were obtained by performing three independent in vitro experiments.

NOTCH3 ^ECD degradation assay

Stably transduced HEK 293T cells were seeded in 6-well plates. When the cell density reached 80%, the cells were treated with 150μg/mL of cycloheximide (CHX, Sigma-Aldrich, St. Louis, MO, USA), a protein synthesis inhibitor, to follow protein degradation over time. At 0, 3, 6, and 9 h after treatment, the cells were collected and cell lysates were prepared for immunoblotting of NOTCH3^ECD using β-actin as a loading control. Data were obtained by performing three independent in vitro experiments.

Mitochondrial transmembrane potential measurement

Mitochondrial membrane potential assay kit with JC-1 (Beyotime, Shanghai, China) was used to analyze mitochondrial membrane potential. Stably transduced HEK 293T cells were seeded in 6-well plates. Next day, after washing once with PBS, 1 mL of cell culture medium and 1 mL of JC-1 working solution were added to the cells. The cells were incubated at 37°C for 30 min. The cells were washed twice with JC-1 (1×), and added 1 mL PBS into each well and blow gently to single cells. Changes in the ΔΨm were analyzed using flow cytometry with flow cytometer (FACS Aria III, BD, USA). The PE and FITC channels were used to detect red JC-1 aggregates in healthy cells and green JC-1 monomers in apoptotic cells. The ΔΨm was represented as the number ratio of the healthy cells with mitochondria that contain JC-1 aggregates to the apoptotic cells with JC-1 monomers. Data were obtained by performing three independent in vitro experiments.

Statistical analysis

SPSS version 22 (IBM, Armonk, NY) was used to conduct statistical analyses. Data were compared by one-way analysis of variance (ANOVA) followed by Dunnett’s post hoc test and expressed as the mean±standard deviation (SD). A value was considered statistically significant at p < 0.05. GraphPad Prism 8.0 software was used for statistical analysis and graphing.

Clinical features of two pedigrees

The proband (Fig. 1a, II-5) of F1, a 45-year-old woman with a previous personal history of hypertension and diabetes, was referred to the neurology department because of recurrent ischemic strokes. She experienced her first ischemic stroke at the age of 37 years, which recurred six times. Gradually, she developed gait impairment, aphasia, cognitive impairment, and apathy. Fluid-attenuated inversion recovery (FLAIR) (MRI 3.0T) revealed symmetric confluent WMH, particularly in the temporal pole and periventricular regions, and multiple LI in the periventricular regions (Fig. 1d). Her father developed memory loss at the age of 55 years and died at the age of 60 years (the cause of death was unknown). Sibling II.2 experienced the first ischemic stroke at the age of 40 years with slurred speech and hemiplegia in early stages, subsequent cognitive impairment and gait difficulties. Other family members were in healthy.

The proband (Fig. 1a, II.1) of F2 was referred to the neurology department owing to memory deficits and progressive gait impairment. His first symptom of migraines without aura occurred at the age of 40 years. The proband experienced his first ischemic stroke resulting in dizziness at the age of 66 years, which recurred twice. Gradually, this developed into memory deficits, gait impairment, irritability, and urinary incontinence. Eventually, he finally died from lung disease at the age of 76 years. FLAIR (MRI 3.0T) included WMH of the external capsule and periventricular regions, multiple LI and cerebral microbleeds (Fig. 1d). III.2 was the 45-year-old son of proband F2 who presented with occasional dizziness and enlarged perivascular spaces in basal ganglia on MRI (Fig. 1d).

GOM deposits were observed by analyzing at least six arteriolar vessels in the 2 probands (Fig. 1c). In addition, the atrophy of VSMCs, stenosis of vascular lumen and thickening of vascular walls were observed. The common clinical traits of the two probands are as follows: 1) all conditions developed in middle age, 2) most patients presented with recurrent ischemic strokes and cognitive impairment, 3) FLAIR revealed WMH in the temporal pole or the external capsule, and 4) GOM deposits were observed on skin biopsy.

NOTCH3 mutation analysis and prediction of bioinformatics

We confirmed a de novo heterozygous missense mutation c.1690G > A in NOTCH3 in the two probands, leading to an amino acid exchange (p. Ala564Thr [p. A564T]) (Fig. 1b). Furthermore, through Sanger sequencing, we observed that II.2, III.10, and III.11 of F1 and III.2 and III.3 of F2 had this mutation. However, we did not observe this mutation in the 120 healthy blood donors.

The NOTCH3 mutation (c.1690G > A, p. Ala564Thr) is encoded by exon 11 and located at the 14th of EGFr domain in NOTCH3^ECD. The minor allele frequency of this mutation was 0.0001 in the gnomAD and ExAC databases and 0 in Asian population. This mutation has not been reported in the HGMD database. Furthermore, the novel mutation was considered to be pathogenic in 18 of the 22 silico prediction tools by VarCards, and the damaging score was 0.82. The closer the damaging score was to one, the more pathogenic the mutation was. Furthermore, Ala at position 564 in the NOTCH3 protein sequence was highly conservative among different species as analyzed using multiple sequence alignments (Fig. 1e). In addition, the protein structure changed from the Ala to the Thr at position 564 in NOTCH3 protein (Fig. 1f).

The clinical features, load of MRI and GOM deposits of patients with p. A564T mutation were similar to patients with cysteine-affecting NOTCH3 mutation. Therefore, we hypothesized that this mutation may also affect the aggregation and degradation of NOTCH3^ECD and ultimately cause cellular dysfunction. Hence, we implemented the following experiments to prove our hypothesis.

The aggregation of NOTCH3 ^ECD A564T mutation in cells

Currently, the pathogenicity of cysteine-sparing NOTCH3 mutations in CADASIL remains controversial. To evaluate the pathogenic role of p. A564T, we successfully established the stably transduced HEK 293T cell line expressing NOTCH3^ECD WT, NOTCH3^ECD A564T mutation and a void vector, and we observed the expression of NOTCH3 in NOTCH3^ECD A564T and WT group at both the protein and mRNA levels, but not in the void vector control (Fig. 2). It was remarkable that the expression of NOTCH3 in NOTCH3^ECD A564T group was higher than that in NOTCH3^ECD WT group and void vector control at both the protein (Fig. 2b) and mRNA levels (Fig. 2c). We have observed this phenotype in cysteine-affecting NOTCH3 mutations [18], and an increased expression of NOTCH3 protein was considered to be a compensatory mechanism for mutation-induced loss of function. Furthermore, we observed the allocation of NOTCH3^ECD in cells using immunofluorescence. We observed that NOTCH3^ECD was not observed in the cytoplasm of void vector control and formed aggregation-like staining in NOTCH3^ECD WT group, however, displayed larger aggregated forms in NOTCH3^ECD A564T group (Fig. 3a). The immunofluorescent density tended to be significantly higher than NOTCH3^ECD WT group and void vector control (Fig. 3b).

OPEN IN VIEWER

Fig. 2

Construction and validation of lentiviral cell lines. a) Protein levels of NOTCH3^ECD are detected by western blot, and β-actin is used as a loading control. b) Quantitative analysis of a. c) Quantitative analysis of the NOTCH3^ECD messenger ribonucleic acid (mRNA) expression. Data are presented as means±SD. ^* p < 0.05, n = 3 per group; one - way ANOVA followed by Dunnett’s post hoc test.

OPEN IN VIEWER

Fig. 3

Aggregation and distribution of NOTCH3^ECD. a) Immunofluorescence staining exhibits NOTCH3^ECD (green) and the nuclei (blue). b) Semi-quantitative analysis of a. Analysis performed using Image Pro Plus software to quantify the mean Integrated Optical Density (IOD) parameter. c) Immunofluorescence staining exhibits Thioflavin-S (Thio-S) positivity (green). d) Semi-quantitative analysis of c. Semi-quantitative analysis of c. Analysis performed using Image Pro Plus software to quantify the mean IOD parameter. e) Double staining exhibits co-localization of aggregated NOTCH3^ECD (red) in Thio-S positive (green) aggregates. Data are presented as means±SD. ^* p < 0.05, n = 3 per group; one - way ANOVA followed by Dunnett’s post hoc test.

We performed Thioflavin-S (Thio-S) staining for evaluating protein aggregation-like in cells. Few Thio-S-positive aggregates were observed in NOTCH3^ECD WT group, void vector control and irrelevant transfected 293T cells, while the larger and more Thio-S-positive aggregates formed in NOTCH3^ECD A564T group (Fig. 3c, d and Supplementary Figure 2). In order to identify aggre-gates in cells, we performed a co-staining usingThio-S and NOTCH3^ECD. The Thio-S-positive agg-regates did not entirely co-located with NOTCH3^ECD in NOTCH3^ECD WT group; however, the co-localization of thio-s-positive aggregates with NOTCH3^ECD was more obvious in NOTCH3^ECD A564T group revealing NOTCH3 A564T mutation caused the formation of aggregation-like (Fig. 3e).

Resistance to the degradation of NOTCH3 ^ECD A564T mutation

The in vitro studies revealed that NOTCH3^ECD A564T mutation was more likely to form aggregates as compared with NOTCH3^ECD WT. Therefore, in order to investigate the possible reasons for higher aggregation of NOTCH3^ECD A564T mutation, CHX pulse-chase experiment was performed. After CHX treatment for 6 h, the NOTCH3^ECD WT cells had considerably degraded. However, the cells harboring A564T mutation remained unchanged. The degradation of NOTCH3^ECD A564T mutation was lower than that of NOTCH3^ECD WT (p < 0.05, Fig. 4). This implied that the NOTCH3^ECD with A564T mutation was more difficult to degrade as compared with NOTCH3^ECD WT.

OPEN IN VIEWER

Fig. 4

Protein stability analysis and degradation of NOTCH3^ECD in cells. a) The protein stability analysis in NOTCH3^ECD WT and NOTCH3^ECD A564T group. The stably transduced HEK 293T cells are chased with 150μg/mL cycloheximide (CHX) at different time points and detected by western blot, and β-actin is used for loading control. b) Quantitative analysis of a, Data are presented as means±SD. ^* p < 0.05 versus the NOTCH3^ECD WT group, n = 3 per group; one - way ANOVA followed by Dunnett’s post hoc test.

Cell viability decreased with NOTCH3 ^ECD A564T mutation

We assessed the effects of the aggregation of NOTCH3^ECD A564T mutation in HEK 293T cells by CCK8 cell viability assays. As presented in Fig. 5a, the viability in NOTCH3^ECD A564T group (77.70±3.43%) significantly decreased as compared with NOTCH3^ECD WT group and void vector control (p < 0.001). However, the viability of cells expressing NOTCH3^ECD WT was not significantly affected as compared with cells expressing a void vector (p = 0.072).

OPEN IN VIEWER

Fig. 5

NOTCH3^ECD A564T mutation - induced apoptosis. a) Quantitative analysis of cell viability in NOTCH3^ECD WT, NOTCH3^ECD A564T group and void vector control. b, d, f) Expression of cleaved caspase-3, caspase-8, BCL2, BAX and cytochrome c (Cyt c) are detected by western blot, and β-actin is used for loading control. c, e, g) Quantitative analysis of b, d, f. Data were presented as means±SD, ^* p < 0.05, n = 3 per group; one - way ANOVA followed by Dunnett’s post hoc test.

Apoptosis increased with NOTCH3 ^ECD A564T mutation

Furthermore, we assessed the potential mechanism of the NOTCH3^ECD A564T mutation, leading to decreased cell viability. Initially, we used western blot to test the apoptosis effector protein caspase-8 and cleaved caspase-3 and we observed that cells overexpressing NOTCH3^ECD A564T mutation exhibited significantly higher expression of caspase-8 and cleaved caspase-3 than that exhibited by cells expressing NOTCH3^ECD WT and a void vector (Fig. 5b, c). Apoptosis can be initiated by caspase-8, which can activate the mitochondria-mediated caspase3-dependent apoptosis pathway [19]. The BAX and BCL-2 located on the outer mitochondrial membrane are involved in the regulation of cell apoptosis; therefore, we further observed the change in BAX and BCL-2. The results revealed that cells overexpressing NOTCH3^ECD A564T mutation exhibited a significantly higher expression of BAX and lower expression of BCL-2 than that exhibited by cells expressing NOTCH3^ECD WT and a void vector (Fig. 5d, e).

Mitochondrial impairment in cells with NOTCH3 ^ECD A564T mutation

Previous studies have demonstrated that upregulation of BAX and downregulation of BCL-2 may affect the changes of ΔΨm and mitochondrial permeability transition pore [20]. We further investigated mitochondrial function by detecting ΔΨm in the stably transduced HEK 293T cells. As indicated in Fig. 6a, the P2 gate represents the healthy cells with mitochondria that contain JC-1 aggregates and the P3 gate represents the apoptotic cells with JC-1 monomers. The number ratio of the healthy cells with mitochondria that contain JC-1 aggregates to the apoptotic cells with JC-1 monomers, which represents the change of ΔΨm, in NOTCH3^ECD A564T group was obviously decreased as compared with NOTCH3^ECD WT group and void vector control (p < 0.05) (Fig. 6b). In addition, cytochrome c in cells expressing NOTCH3^ECD A564T mutation was significantly higher than that present in cells expressing NOTCH3^ECD WT and a void vector (p < 0.05); however, cytochrome c in mitochondria was significantly lower (p < 0.05) (Fig. 5f, g). This indicated that the mitochondrial permeability transition pore may be open and cytochrome c in the mitochondria was transported to the cytoplasm, thus activating the caspases-3-dependent apoptotic pathway.

OPEN IN VIEWER

Fig. 6

Analysis of the mitochondrial membrane potential (ΔΨm) in NOTCH3^ECD WT, NOTCH3^ECD A564T group and void vector control by JC-1 staining. a) The P2 gate represents the healthy cells with mitochondria that contain JC-1 aggregates and the P3 gate represents the apoptotic cells with JC-1 monomers. b) The ΔΨm of was represented as the ratio of the number of the healthy cells with mitochondria that contain JC-1 aggregates to the apoptotic cells with JC-1 monomers. ^* p < 0.05; one - way ANOVA followed by Dunnett’s post hoc test.